The Interest Rate Risk of Mortgage Loan Portfolio of Banks

A Case Study of the Hong Kong Market

Jim WongHong Kong Monetary Authority

Paper presented at the Expert Forum on Advanced Techniques on Stress Testing: Applications for SupervisorsHosted by the International Monetary FundWashington, DC– May 2-3, 2006

The views expressed in this paper are those of the author(s) only, and the presence of them, or of links to them, on the IMF website does not imply that the IMF, its Executive Board, or its management endorses or shares the views expressed in the paper.

The Interest Rate Risk of Mortgage Loan Portfolio of Banks

A Case Study of the Hong Kong Market

Jim WongMarket Research Division

Hong Kong Monetary AuthorityMay 2006

3

Content

• Market and Interest Rate Structures

• Interest Rate Risk

• Stress-testing Framework and Model

• Scenario Analysis

4

Mortgage Market in Hong Kong

• Primarily adjustable rate mortgages, with pricing mainly referenced to prime rate (P)

• The cost of fund is determined by a mix of interest rates

• Dominated by major retail banks- not only by market shares- also by price setting

5

Interest Rate Structure of Hong Kong

• Under the Linked Exchange Rate system, interest rates in Hong Kong tend to track closely their US counterparts

• Major local interest rates are:- Prime rate (P, this is also named as the best lending rate)- Savings deposit rate (SR)- Interbank rates (HIBORs)- Time deposit rates (TDRs)- Others (such as interest rates on negotiable certificates ofdeposit and debt instruments)

6

Prime rate• Set by major retail banks and followed by other banks• normally adjusts when there is a change in the US Federal funds

target rate (FFTR), or when pressures from changes in the cost of funds are built up to a certain level

0

2

4

6

8

10

12

Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06

% p.a.

Average P3-month LIBOR

FFTR

7

Savings deposit rate• Normally adjusts in tandem with P• Traditional practice, not required by regulation

0

2

4

6

8

10

12

Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06

% p.a.

Average P

Average SR

8

Interbank rates• Normally track closely US interest rates, but may deviate from the

US rates with a risk premium due to liquidity conditions and other factors

0

2

4

6

8

10

12

Jan-89 Jan-91 Jan-93 Jan-95 Jan-97 Jan-99 Jan-01 Jan-03 Jan-05

3-month HIBOR3-month LIBOR

% p.a.

9

Risk premium• HIBOR minus LIBOR• Short-term deviations in interest rates of the Hong Kong dollar

over the US dollar. The mean level over the past 17 years is close to zero

-300

-200

-100

0

100

200

300

400

500

600

700

Jan-89 Jan-91 Jan-93 Jan-95 Jan-97 Jan-99 Jan-01 Jan-03 Jan-05

Bps

Historical mean95% confidence band

Risk Premium

10

Reasons for the narrowing of spread

Balance sheet positions of banksAssets Liabilities

Fixed-rate assets Fixed-rate liabilities

Variable-rate assets (mostly HIBOR-based lending)

Variable-rate liabilities (mostly HIBOR-based borrowings)

P-based assets (largely priced with reference to P)

P-based liabilities (largely savings deposits, the interest rates of which move in tandem with the P)

Non-interest bearing assets (such as equipment and buildings)

Non-interest bearing liabilities (mainly demand deposits and capital)

11

0

2

4

6

8

10

12

Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05

Average P

Weighted average mortgage rate

% p.a.

P + 1%

P - 3%

12

• This was made possible by the extraordinary low funding cost with the HIBOR being at an usually deep discount to LIBOR of 200 bps.

0

2

4

6

8

10

12

Jan-89 Jan-91 Jan-93 Jan-95 Jan-97 Jan-99 Jan-01 Jan-03 Jan-05

3-month HIBOR

3-month LIBOR

% p.a.

200 bps

13

Stress-testing framework and model

• An error correction model is established to examine how P and other interest rates may move in response to changes in FFTR (and LIBOR) and risk premium.

• With the estimation results, we could stress test and assess the potential impact on interest margin of banks’existing mortgage portfolio under an interest rate shock of a possible sharp reversal of the risk premium.

14

Stress-testing framework and modelModel specification:

ttttt eDSXRP ++++=∆ − 12110 ααα (1) where tX is

1 ratioL/A based-P

with nsObservatio

1 ratioL/A based-P

with nsObservatio

when

08

0743

06

0521

<

≥

⎪⎪

⎩

⎪⎪

⎨

⎧

∆+∆+∆+∆

∆+∆+∆+∆

=

∑∑

∑∑

=

−−

=

+−

−+

=

−−

=

+−

−+

M

llt

M

llttt

M

llt

M

llttt

t

prempremFFTRFFTR

prempremFFTRFFTR

X

φφφφ

φφφφ

(2) with

ttt FFTRPR 1β−= (3)

15

Stress-testing framework and model

•The error correction model studies the situations (total 4 situations) :

– FFTR/Risk Premium are increasing or decreasing

– P-based L/A ratio is greater than or less than 1

16

Stress-testing framework and modelEstimation Results

Dependent Variable : P Jan 95-Jan 05

Selected Variable Coefficient

Short run (1) A

tR 1− -0.09 * L/A ratio ≥ 1

+∑∆t

prem 0.12 ** −∑∆t

prem

0.54 ** +∆ tFFTR 0.16 * −∆ tFFTR 0.74 ***

Long run (2) FFTR 0.89 ***

Adj. R-squared 0.69 Q(6) 11.75 Q(12) 17.83 DW 1.82

17

Scenario analysisThe scenario that the risk premium increases by 200 bps (back to its historical mean level), while US interest rates remain unchanged, is examined. The effects in three months’ time are as below:

Risk Premium Reversal of 200 bps (back to the mean level)

Changes (bps) P +24 HIBOR +200 Interest Margin (P – HIBOR) -176

18

Scenario analysisThe impacts on net interest margin of HIBOR-based mortgage loans priced in January 2005 are simulated.

Risk Premium Reversal of 200 bps (back to the mean level)

Mortgage Pricing 200 Funding Cost -65 Current Gross Mortgage Margin 135 Operating Cost -30 Credit Cost -10 Current Net Mortgage Margin 95 Estimated Reduction of Mortgage

Margin -176

Simulated Net Mortgage Margin After Impact -81